Ambrose Bonnaire-Sergeant

Another insight is that if all `:lets` and `:letks` bindings are never used in the body, we can treat it like a static context. This is because the bindings are...

Yes, if I can't do enough dynaload magic to make rmf optional I will make a new namespace for muuntaja or sequester the rmf stuff in another helper namespace. I'm...

I'm reconsidering whether I need to add rmf support to 2.x. I figured out how to [support 1.x `fn?` coercions in 2.x](https://github.com/metosin/compojure-api/commit/113263b05c4ecb5b63fe187a2e1830377d9b54b5), perhaps the remaining differences aren't as important.

IMO it's a bit inconsistent with the rest of malli that the bounds are clamped on behalf of the user in the generator. Otherwise the rest looks good.

Are there other examples of truncation in malli that you are following? To my taste this should mirror bounding a double with a bigint, what happens there? Or a hypothetical...

> Based on your notes, I think we can agree that we should throw when the user provides boundaries that are unsupported by 32-bit floats. Can we also agree that...

IMO this is useful. Perhaps a simpler cross-platform implementation would be `m/eval` `:gen/gen` if it is not a generator, and then have users of the schema manually load the generator...

@ikitommi I took your idea and ran with it. Since you were unhappy with `:function`, I added a new proxy schema `:ifn`. LMK what you think. ```clojure [:function [:=> [:cat]...

> also, `:ifn` sound bit like `ifn?` which is a different thing, e.g. a map gives true to that. I don't see the problem, that's also true for `:function`. ```clojure...

FWIW this regal schema might also be useful, I'm not sure how a regal generator and test.chuck generator compare though: https://github.com/lambdaisland/regal/pull/48